Toby Melville / Reuters

It took, at most, several seconds. An enormous hunk of rock, roughly the size of Manhattan, came whirling out of the vastness of space. It pierced Earth’s thin atmosphere, ignited as it fell, and slammed into the crust, opening a crater 20 miles deep in modern-day Mexico’s Yucatan Peninsula.

Of course, it killed the non-avian dinosaurs: How could it not? By its end, the cataclysm wiped out 75 percent of all species that dwelled on Earth. In the last quarter century, we have gotten used to seeing images of that catastrophe: of the hellfire that rained down to Earth, igniting massive forest fires; of the years-long “impact winter” that dimmed the sun and chilled the Earth.

But less well-known is what followed that winter. Scientists believe that the asteroid, which struck Earth roughly 66 million years ago, eventually triggered a lengthy period of ferocious global warming. Upon impact, it vaporized solid limestone into gas, and it incinerated enormous swaths of forest. This unleashed so much heat-trapping carbon dioxide into the atmosphere so quickly that, across all of Earth’s history, its rate of increase seems to be rivaled only by recent carbon pollution from factories, cars, planes, and modern industry.

A study published Thursday in Science finds new evidence of that warming while setting it in a dreadful context. It may have taken seconds for the asteroid to chew a 20-mile-deep hole in Earth. But, its authors say, it took roughly 100,000 years for Earth’s climate to return to normal.

The research argues that Earth’s average temperature was elevated by 5 degrees Celsius for the 100 millennia that followed the impact. Notably, it supports this assertion not just with computer models, but with direct, observed evidence from the time period. By analyzing the bones of fish that lived in modern-day Tunisia before, during, and after the impact, scientists were able to detect a planet-sweltering warming signal.

“We think we’ve found the best evidence yet—and pretty good, rigorously constrained evidence—for warming,” said Kenneth MacLeod, one of the authors of the study and a professor of geology at the University of Missouri. “I think it’s a big step forward in documentation of the temperature history of this interval.” The massive warming that followed the asteroid’s impact issues a clear warning for today’s warming world, MacLeod told me.

What can fish bones tell us about the climate? A surprising amount. MacLeod said they act as a “paleothermometer,” preserving a prehistoric temperature even after tens of millions of years of exposure to other disturbances. That’s not because of their fishiness, necessarily, but because of the oxygen atoms preserved in their hard tissue. When oxygen “mineralizes”—that is, when it’s removed from water to form teeth or bone—it does so in a specific way. Heavier oxygen atoms, with 18 neutrons in their nucleus, are more likely to enter the new bone than lighter oxygen atoms, with only 16 neutrons. These heavier atoms are preferred at a specific rate, but that rate decays in a predictable way as it gets warmer, MacLeod said.

“For every five degrees Celsius of temperature increase, there’s one atom of oxygen-18 that stays in the water rather than go into a mineral,” he told me. “With modern mass spectrometers, we can measure the oxygen-16 to oxygen-18 ratio in carbonates [like bone] to less than .1 parts per thousand.” Even after accounting for error, this means they can detect a degree or two of temperature change millions of years after the fact.

David Beerling, a climate scientist at the University of Sheffield who was not connected to the new paper, said that the research was worthwhile. In 2002, Beerling was one of the first scientists to propose that a period of global warming followed the Chicxulub impact. The temperature estimates in that paper, he said, were provisional and based primarily on computer models. “Independent support from ocean-sediment records [for the warming] is welcome and will reinvigorate debate,” he told me in an email.

The confirmation of a 100,000-year warm period also raises a new question: Did global warming, and not merely “the asteroid,” make the dinosaurs go extinct? Previous studies have argued that the last non-avian dinosaurs survived roughly 30,000 years after the Chicxulub impact—which means that at least some of them endured well past the impact winter.

MacLeod isn’t yet ready to say that elevated global temperatures are what finally did the dinosaurs in. He’s currently working with another scholar, he said, to identify “specific mechanisms that were most likely responsible for killing different victims.”

“At the moment, it is safe to say the warming is another insult to those species trying to run the gauntlet between the latest Cretaceous and earliest Paleogene, including any dinosaurs that survived the immediate aftermath of the impact and subsequent impact winter,” he told me in an email.

And there may be more bad news in the study for humanity. While scientists believe that Earth’s atmosphere filled with carbon dioxide after the impact, they’re not sure how much carbon dioxide was actually released. One study finds the amount of carbon dioxide in the atmosphere increased by fivefold after the impact. “But that’s really big, it’s just one study, and geologists’ ability to estimate CO₂ levels across time is notoriously poor,” said MacLeod.

If that huge, fivefold shift is accurate, then the 5 degrees Celsius of temperature increase detected by MacLeod and his colleagues in the fish bones makes sense. The best modern climate models predict that that much warming would follow from such a gigantic release of carbon dioxide.

Yet many more studies do not argue that so much carbon dioxide was released. They say the asteroid may have only doubled carbon levels in the atmosphere. “If the increase was actually by a factor of two, or a factor of three, and if our estimate of five degrees Celsius of warming is representative of the whole globe —then the climate models are underestimating how much it will warm,” MacLeod told me.

In other words, modern climate models may be underestimating how much warming the current carbon-dioxide increase would produce. And that would be bad: It could suggest that we could see as much as twice the warming as the Intergovernmental Panel on Climate Change currently forecasts for the end of the century.

“Models are tested by their ability to reproduce what we’ve seen—and we’ve never seen anything like CO₂ levels changing at the rate they have over the last 100 years,” said MacLeod. “Or at least, we haven’t seen them since 66 million years ago.”

But that revelation lies further off. For now, the 100,000-year warmed period—produced by just the carbon emissions that occurred in the span of a human lifetime—are worrying enough. “The obvious parallel is that what we’ve done in the last century will take another 100,000 years to decay to its pre-perturbation state,” said MacLeod. “Simply stopping emissions won’t get you to where you were before, because that carbon dioxide takes a long time to cycle through our system.”

Sounds messy.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.